Spark on YARN 环境搭建

1. 修改spark-env.sh

cd /export/server/spark/conf cp spark-env.sh.template spark-env.sh vim /export/server/spark/conf/spark-env.sh 添加以下内容: HADOOP_CONF_DIR=/export/server/hadoop/etc/hadoop YARN_CONF_DIR=/export/server/hadoop/etc/hadoop SPARK_HISTORY_OPTS="-Dspark.history.fs.logDirectory=hdfs://node1:8020/sparklog/ -Dspark.history.fs.cleaner.enabled=true" 同步到其他两台(单节点时候, 不需要处理) cd /export/server/spark/conf scp -r spark-env.sh node2:$PWD scp -r spark-env.sh node3:$PW

2.修改hadoop的yarn-site.xml

cd /export/server/hadoop-3.3.0/etc/hadoop/ vim /export/server/hadoop-3.3.0/etc/hadoop/yarn-site.xml 添加以下内容:                 yarn.resourcemanager.hostname         node1                 yarn.nodemanager.aux-services         mapreduce_shuffle                     yarn.nodemanager.resource.memory-mb         20480                 yarn.scheduler.minimum-allocation-mb         2048                 yarn.nodemanager.vmem-pmem-ratio         2.1                      yarn.log-aggregation-enable         true                     yarn.log-aggregation.retain-seconds         604800                     yarn.log.server.url         http://node1:19888/jobhistory/logs                     yarn.nodemanager.pmem-check-enabled         false                 yarn.nodemanager.vmem-check-enabled         false     将其同步到其他两台 cd /export/server/hadoop/etc/hadoop scp -r yarn-site.xml node2:$PWD scp -r yarn-site.xml node3:$PWD 3.Spark设置历史服务地址 cd /export/server/spark/conf cp spark-defaults.conf.template spark-defaults.conf vim spark-defaults.conf 添加以下内容: spark.eventLog.enabled                  true spark.eventLog.dir                      hdfs://node1:8020/sparklog/ spark.eventLog.compress                 true spark.yarn.historyServer.address        node1:18080 配置后, 需要在HDFS上创建 sparklog目录 hdfs dfs -mkdir -p /sparklog 4.设置日志级别: cd /export/server/spark/conf cp log4j.properties.template log4j.properties vim log4j.properties 修改以下内容: log4j.rootCategory=WARN, console 同步到其他节点 cd /export/server/spark/conf scp -r spark-defaults.conf log4j.properties node2:$PWD scp -r spark-defaults.conf log4j.properties node3:$PWD 5.配置依赖spark jar包 当Spark Application应用提交运行在YARN上时，默认情况下，每次提交应用都需要将依赖Spark相关jar包上传到YARN 集群中，为了节省提交时间和存储空间，将Spark相关jar包上传到HDFS目录中，设置属性告知Spark Application应用。

hadoop fs -mkdir -p /spark/jars/ hadoop fs -put /export/server/spark/jars/* /spark/jars/ 修改spark-defaults.conf cd /export/server/spark/conf vim spark-defaults.conf 添加以下内容: spark.yarn.jars  hdfs://node1:8020/spark/jars/*

同步到其他节点(无需分发, spark只有一个单节点)

cd /export/server/spark/conf scp -r spark-defaults.conf root@node2:$PWD scp -r spark-defaults.conf root@node3:$PWD

6.启动服务

Spark Application运行在YARN上时，上述配置完成

启动服务：HDFS、YARN、MRHistoryServer和Spark HistoryServer，命令如下：

## 启动HDFS和YARN服务，在node1执行命令 start-dfs.sh start-yarn.sh 或 start-all.sh 注意：在onyarn模式下不需要启动start-all.sh（jps查看一下看到worker和master） ## 启动MRHistoryServer服务，在node1执行命令 mr-jobhistory-daemon.sh start historyserver ## 启动Spark HistoryServer服务，，在node1执行命令 /export/server/spark/sbin/start-history-server.sh

1. Spark HistoryServer服务WEB UI页面地址：http://node1:18080/
2. 测试代码

1，spark 自带sample

cd /export/server/spark/bin

./spark-submit \
--master yarn \
--conf "spark.pyspark.driver.python=/root/anaconda3/bin/python3" \
--conf "spark.pyspark.python=/root/anaconda3/bin/python3" \
/export/server/spark/examples/src/main/python/pi.py \
10

http://node1:8088/

​​​​​​​